in regression model control can be taken as beta zero

You can just include control variables into the same model.

Dear Prof. Necmettin, please guide the way through which they can be used.

You should be more explicit about your aim. If you want to control for the effects of some variables on some dependent variable, you just include them into the model. Say, you make a regression with a dependent variable y and independent variable x. You think that z has also influence on y too and you want to control for this influence. Then you add z into the model as a predictor (independent variable).

Dear Prof Mehmet, I want to control the influence of these variables.

So, i can include the variables which may have influences over the DV along with the other IVs and run regression excluding the control variables, is it?

After including the Control variable in IV list how do you interpret output that these are controlled.

Amina Charania See what happened with the original relationship? The R2, the coefficient , the coefficient sign, significance, are they changed or still the the same . If there is change, its mean that there is controlling effect. The difference in these parameter comes from the Control variable. If you want to interpret it then Interpret it as like other IV. As with one unit of change in control variable this much change will occur in DV keeping the other IVs constant.

The main conclusion for a case where the significance is still there, it shows a relationship between IV and DV remains. This does however not imply that we now have showed that there is a causal effect. There is still a lot of other relevant variables to control for, and in a thesis you should definitely do. But the principle is the same, we would only add more variables to the regression analysis.

no

The purpose of a control variable is to isolate the effect of the explanatory variable(s) on the outcome variable. Therefore, a control variable ought to be an obvious determinant of the dependent variable. Before regressing the DV on the IVs, check for the effect of the control variable. Later, pool the controls and the IVs in the same regression model.

Regards

Mr Mehmet Sinay,

But won’t the R and R square account for the total significance of all variables then. How do I distinguish the main variable. Say I want to do a regression how does corruption affect the number of equity shares per household in a given country. Do I need to account for things such as GDP, inflation or interest rates? or some other things?